Problem with latest APBS version and make_apbs_inputs

[Lane Votapka]

Hi Gary,

I think there might be a problem with the compatibility between the latest versions of APBS and make_apbs_inputs. It seems as though the latest versions of APBS add a pesky "-PE0" within the DX file name.

For instance, if I have a file called "apbs_input.xml" with the following lines:

"""

<group>
         <name>receptor</name>
         <core>
            <name>receptor</name>
            <atoms>trypsin.xml</atoms>
            <all_in_surface>false</all_in_surface>
            <is_protein>false</is_protein>
            <dielectric>4.0</dielectric>
            <grid_spacing>0.5</grid_spacing>
         </core>
      </group>

"""

And then run:

make_apbs_inputs apbs_input.xml > input.xml

The file input.xml will have the following lines:

"""

<group>    
      <name> receptor </name>
      <core>      
        <name> receptor </name>
        <atoms> trypsin.xml </atoms>
        <all_in_surface> false </all_in_surface>
        <is_protein> false </is_protein>
        <dielectric> 4.0 </dielectric>
        <grid_spacing> 0.5 </grid_spacing>
        <electric_field checked="true">        
          <grid source="make_apbs_inputs"> receptor0.dx </grid>
          <grid source="make_apbs_inputs"> receptor1.dx </grid>
        </electric_field>
      </core>
    </group>

"""

Yet when I do "run_apbs_inputs" on input.xml, then I get errors, and the following file is generated:

receptor0-PE0.dx

I think there's a mismatch between the file name made by APBS, and what is expected by make_apbs_inputs, at least for these later versions of APBS.

-Lane

--

Lane Votapka Ph.D.

Staff Scientist, Amaro lab

UC San Diego

[Gary Huber]

Hi, Lane,

I thought I had adjusted for that possibility, but I'll double-check.  That PE0 extension does pop up from time to time.

Gary

--
You received this message because you are subscribed to the Google Groups "browndye-users-l" group.
To unsubscribe from this group and stop receiving emails from it, send an email to browndye-users...@ucsd.edu.

[Robert Konecny]

Hi Lane,

The "-PEn" suffix is used only if apbs was compiled with MPI, regardless of
apbs version. Serial apbs doesn't use it. See:

https://urldefense.com/v3/__https://github.com/Electrostatics/apbs/blob/main/src/routines.c*L1990__;Iw!!Mih3wA!DXQ8PJ-w1f1urJPOa4yTwP_uinyy5h9mavaAYQpR17YP5Zlp_krOWiyGVt4AWHRIfOZLPkZqqvh0ieo9Db-k$

Each thread in parallel APBS writes out separate .dx file (or any other
output file) so the -PEn suffix is used to separate them.

Best,

Robert

>--
>You received this message because you are subscribed to the Google Groups
>"browndye-users-l" group.
>To unsubscribe from this group and stop receiving emails from it, send an email

>to [1]browndye-users...@ucsd.edu.
>
>References:
>
>[1] mailto:browndye-users...@ucsd.edu

[Lane Votapka]

Hi All,

Ah yes, that makes sense, Robert.

Also, Gary, I realized what the problem was - I had been encountering this problem when I was running a CI workflow on Github Actions, and realized that, by default, the workflow was downloading and installing a version of Browndye2 from 2021. Silly mistake. The workflow now downloads the latest version of Browndye2 and the problem has disappeared.

Thank you.

-Lane

On Fri, Jul 28, 2023 at 10:48 AM Robert Konecny <r...@ucsd.edu> wrote:

Hi Lane,

The "-PEn" suffix is used only if apbs was compiled with MPI, regardless of
apbs version. Serial apbs doesn't use it. See:

https://github.com/Electrostatics/apbs/blob/main/src/routines.c#L1990